Metal–Polymer–Metal Split‐Dipole Nanoantennas

Abstract

The conjugated polymer semiconductor poly(3-hexylthiophene), (P3HT), is integrated directly into the slot region of resonant plasmonic split-dipole nanoantennas. The P3HT radiative emission rate is enhanced by a factor of up to 29, in experiment, and 550 for the ideal case, due to the large local density of optical states in the nanoantenna slot region. Additionally, the theoretical modified luminescence quantum efficiency is shown to increase from 1% to 45% for optimized nanoantenna parameters. The ability to modify semiconductor optical properties using nanostructures that support surface plasmon resonances has enabled the demonstration of improved light-management in conventional opto-electronic devices, such as light-emitting diodes,1-3 lasers,4 and photovoltaic devices,5, 6 as well as in integrated optical computing and light localization on the nanoscale.7-11 In particular, resonant plasmonic nanoantennas can be used to control the polarization, directivity, light-emission intensity and decay rate of coupled emitters via the large local density of optical states in the nanoantenna's near-field.7-16 Recent theoretical studies have proposed linear metal–semiconductor–metal nanoantennas where semiconductor material is incorporated into a central ‘slot region’ of a noble metal nanorod, in analogy to the feed element in radio frequency antennas.11, 17 This not only allows the full benefit of local-field effects in the antenna near-field to be exploited, but also permits light to be efficiently in/out coupled to sub-wavelength semiconductor material volumes. However, realizing integrated metal–semiconductor nanoantenna structures such as these remains challenging. Additionally, the application of plasmonic nanoantennas to organic semiconductor materials for improved light management has been largely unexplored despite the potential for enhancing emission rate and quantum efficiency. In particular, organic conjugated polymer semiconductors such as polythiophene, which exhibit high carrier mobilities but possess relatively poor luminescence properties,18-20 would benefit from plasmonic nanoantennas, potentially opening up opportunities for use as the active material in organic light-emitting opto-electronic devices. Here, we integrate the conjugated polymer semiconductor poly(3-hexylthiophene), P3HT, directly into the slot region of resonant plasmonic split-dipole nanoantennas. Using this approach, P3HT radiative emission rate is enhanced by a factor of up to 29, in experiment, and 550 for the ideal case (theory). Additionally, theoretical modified luminescence quantum efficiency is shown to increase from 1% up to 45% for optimized nanoantenna parameters. This work demonstrates that integrated metal–polymer–metal nanoantennas could enable a new generation of high-performance conjugated polymer optoelectronic devices. Theoretical design of resonant gold–polythiophene–gold split-dipole nanoantennas. a) Top left: Schematic of split-dipole plasmonic nanoantenna with P3HT in the slot region. Bottom left: Molecular structure of P3HT. Right panels: Theoretical 2D z-polarized electric-field intensity cross-sections through nanoantennas with different lengths. b) Theoretical radiative and non-radiative decay rate enhancement (relative to a dipole in P3HT in the absence of a nanoantenna) and modified quantum efficiency for a dipole emitter placed in the center of the P3HT slot region as a function of L (dipole oriented parallel to the nanoantenna long axis). c) Theoretical radiative decay rate and d) quantum efficiency enhancements as a function of L and d in the half-wave nanoantenna regime. All theoretical data was simulated using 3D full-field FDTD software using a free-space wavelength of 700 nm (see Experimental Section). where is the radiative decay rate in P3HT without the gold nanoantenna and is the intrinsic luminescence quantum efficiency of P3HT (∼1%)19, 20 which accounts for the large intrinsic non-radiative decay rate, , associated with P3HT (i.e., ); see SI for derivation of Equation 1). Figure 1b is a plot of the enhancement in and (relative to a dipole emitter in P3HT without the metal nanoantenna) and for a range of nanoantenna lengths. While was relatively unchanged with nanoantenna length, and were enhanced by factors of 53 and 25, respectively, at the half-wave resonance due to an efficiently radiating nanoantenna and to a lesser degree at higher-order resonances. At optical frequencies, it is expected that the radiative efficiency of a nanoantenna depends not only on its length but also on its diameter.17 Figure 1c is a 3D plot of as function of d and L around the half-wave resonance of the nanoantenna (here nbk = 1). Optimum enhancement in , by a factor of 550, occurred for d = 27 nm and L = 180 nm and values of up to 45% were achievable for diameters in the range 28 to 45 nm and nanoantenna lengths between 180 and 230 nm. The diameter and length at which was a maximum were larger than those for due to a greater contribution of non-radiative decay rate enhancement for smaller diameters (Figure 1d). Therefore, nanoantennas that are designed to optimize quantum efficiency require slightly larger dimensions than antennas that are designed to optimize radiative decay rate. Although the simulations were carried out for a single dipole source placed in the center of the P3HT slot region, large enhancements were expected for dipoles positioned at other locations within the slot (SI, Figure S3). Therefore, in fabricated structures an ensemble of P3HT emitters in the slot region are expected to be modified by the nanoantenna. Split-dipole nanoantenna heterostructures were fabricated using a template-directed sequential electrodeposition process followed by a metal evaporation step (Figure 2 and Experimental Section). The template employed was nanoporous alumina which was grown by anodizing 300–600 nm of aluminum on a conductive substrate (ITO or gold) in oxalic acid and subsequent pore widening (Figure 2a,b).23-25 Gold and P3HT were electrodeposited into the alumina template pores in sequence under pulsed potentiostatic conditions (Figure 2c,e). The height of the electrodeposited gold segments was varied between 20 and 150 nm across a given sample and the height of P3HT deposited on the gold segment was kept constant (typically at a value between 20 and 50 nm depending on the number of applied pulses). To complete the split-dipole nanoantenna structure, 40–80 nm of gold was evaporated through the alumina template onto the ends of the P3HT segments (Figure 2d). To ensure the gold was deposited into the template pores during this final step it was critical for the thickness of the template to be ≤600 nm. Multiple regions with nanoantennas of different lengths were fabricated on a single substrate by varying the deposition time of the gold segments within each region (Figure 2f). Schematics of the fabrication process for gold–P3HT–gold split-dipole nanoantenna heterostructures by sequential electrodeposition and thermal evaporation in nanoporous alumina templates on conductive substrates: a,b) 300–600 nm of aluminum on gold or ITO was anodized in oxalic acid at 35 V to form a supported, electrically contacted, nanoporous alumina template. Alumina grown in this way had cylindrical pores (55 nm in diameter) which were vertically oriented relative to the conductive substrate. c) Following a pore widening step and alumina barrier etch, gold and P3HT were electrodeposited within the template pores, sequentially. d) Nanoantenna fabrication was completed by thermally evaporating 40–80 nm of gold through the pores of the alumina template onto the P3HT segments. e) Typical current density versus time plots acquired during pulsed electrodeposition of gold and P3HT in nanoporous alumina templates. For gold deposition, -3 V pulses (versus a platinum mesh quasi-reference electrode, QRE) of 0.02 s duration were applied to the working electrode with -0.2 V applied for 5 s between pulses. For P3HT electrodeposition (electropolymerization from the 3HT monomer), +3 V pulses (versus Pt QRE) of 0.02 s duration were applied with -0.6 V for 5 s between pulses. f) A true-color photograph of a sample with twelve different regions with nanoantennas of different lengths (acquired at an angle of ∼50° off-normal to the substrate plane under halogen lamp lighting; regions are delineated by white dashed lines). Figure 3a and b show cross-sectional scanning electron microscopy (SEM) images of gold–P3HT–gold split-dipole nanoantennas after fabrication in an alumina template and after template removal, respectively. Typically, the fabricated nanoantennas exhibited rounded or cone-shaped gold segments on the top side due to shadowing by the template pore walls during the metal evaporation step (Figure 3a–d). Simulations of asymmetric split-dipole nanoantennas with one cone-shaped gold segment showed that, compared to symmetric nanoantennas with similar total length, the resonance wavelength of the nanoantenna did not change significantly, though electric-field intensity in the slot region decreased by a factor of ∼2 (SI, Figure S4). Cross-sectional SEM images of vertically oriented gold–P3HT–gold nanoantenna arrays a) in the alumina template and b) after alumina template removal (acquired at a 45° sample tilt). c,d): SEM images of single nanoantennas with different segment lengths (on gold and silicon substrates, respectively). e) Reflectance spectra of a vertically oriented array of nanoantennas with L = 196 ± 21 nm acquired using a 0.2 n.a. objective and a 0.65 n.a. objective (relative to planar aluminum). f) Differential extinction spectra for arrays of nanoantennas with lengths of 47 ± 7 nm and 196 ± 21 nm (calculated from the reflectance spectra as [(1–R0.65)–(1–R0.2)]/(1–R0.2), where R0.65 and R0.2 are the reflection intensities from the nanoantenna arrays for the 0.65 n.a. and 0.2 n.a. objectives, respectively, relative to planar aluminum). Inset: Scattered-light spectrum of a single split-dipole nanoantenna (L ∼ 190 nm) embedded in optical epoxy with refractive index of 1.48. To detect longitudinal resonances of nanoantennas vertically oriented on the growth substrate, reflection spectra of arrays with different lengths were collected with objectives of different numerical aperture, n.a. (Figure 3e). The reflection spectrum of an array of nanoantennas with L = 196 ± 20 nm collected with a 0.2 n.a. objective showed a broad minimum at 510 nm, which was attributed to transverse nanoantenna resonances. Reflection spectra acquired with a 0.65 n.a. objective allowed longitudinal antenna resonances to be more effectively excited and collected and, as a result, an additional minimum was apparent in the reflected-light spectrum at ∼670 nm. The differential extinction in an angular range of 11° to 35° off-normal incidence exhibited a pronounced peak at 690 nm (Figure 3f). In contrast, the differential extinction spectrum of a nanoantenna array with L = 47 ± 7 nm did not show longitudinal resonances in the visible range of the spectrum. The scattered-light spectra from isolated nanoantennas with lengths in the 100–250 nm range exhibited narrow, pronounced resonances with peak maxima in the 600 to 800 nm wavelength range (Figure 3f, inset). Resonant enhancement of P3HT emission from gold–P3HT–gold split-dipole nanoantennas. a) PL intensity enhancement () spectra (100× oil immersion objective (n.a. = 1.3)) for nanoantenna arrays with L values of 47 ± 7 nm, 82 ± 8 nm, 116 ± 16 nm, and 130 ± 13 nm. Inset in (a): Plot of versus L at a wavelength of 700 nm (4 spectra per L value). b) PL lifetime decays acquired from arrays with average L of 116 nm and 47 nm (collected in the 650 to 750 nm wavelength range), along with a PL lifetime decay from a GaAs nanowire, representing the response of the measurement system. Inset in (b): Total PL lifetime () as a function of L (4 measurements per L value). c) PL lifetime decays acquired from a P3HT thin film (green), a neat P3HT nanowire (blue), a single resonant gold-P3HT monomer nanoantenna heterostructure (wine) and a single resonant gold-P3HT-gold split-dipole nanoantenna heterostructure (red). The instrument response function is shown in grey. Solid black lines are exponential fits to the data–the thin film data was fit with a single exponential and all others were fit with a double exponential function. d) Radiative and e) non-radiative decay rates calculated from and as a function of nanoantenna length (both and were determined from PL lifetime decays; wavelength range 650–750 nm). The dashed grey lines represent the radiative and non-radiative decay rates of a neat P3HT nanowire. A double-Gaussian function was fit to the data in (d) (black dashed curves). where (see SI). Here, was determined from the PL lifetime decays at time = 0 ps for the wavelength range 650–750 nm. For calculation of it was assumed that: 1) excitation rate enhancement was negligible (i.e., is well away from local surface plasmon resonances of the nanoantennas); and 2) PL intensity enhancement was a measure of quantum efficiency enhancement, i.e., .12, 15, 26, 27 Note that , the radiative decay rate of P3HT in the absence of a resonant nanoantenna was determined from the total PL lifetime value for neat P3HT nanowires according to the expression . was estimated to be 0.014 × 109 s−1; taking = 1%. Values for were calculated from Equation 2 and plots of and versus L are shown in Figure 4d and Figure 4e, respectively (noting that ). exhibited clear changes with L, reaching values greater than 0.4 × 109 s−1 at L = 116 nm and 0.15 × 109 s−1 at L = 196 nm (enhancements of 29 and 11, respectively, relative to neat P3HT nanowires). of was largest up to 16 × 109 s−1 for short nanoantenna lengths (L = 76 nm) and decreased to ∼6 × 109 s−1 for larger L (enhancement in of ∼4 relative to that of neat P3HT). These experimentally derived radiative and non-radiative decay rates are in qualitative agreement with those predicted with simulation in Figure 1, with the radiative decay rate component enhanced periodically with increasing length, and the non-radiative component remaining relatively unchanged. However, both the radiative decay rate enhancement values and the resonance length determined from experiment are significantly smaller than the theoretical values. Many experimental factors are likely to have contributed to these differences including: 1) non-optimal P3HT dipole orientation and variation in the local density of optical states at various locations within the P3HT slot region; 2) deviations in nanoantenna shape from the ideal case (see SI, Figure S4); 3) substrate effects which are likely to increase the effective index of the nanoantennas’ environment; 4) bulk array effects (nanoantenna separation distances of between 20 to 60 nm in the arrays are likely to cause a degree of coupling or interference between adjacent nanoantennas); and 5) variation in . Further investigation of experimental deviations from theory are outside the scope of the current paper but will be considered in more detail in later work on nanoantenna/conjugated polymer hybrid systems. While ∼ 1% for P3HT, it is important to consider how such a nanoantenna geometry could be applied to other organic semiconductor materials with different values. The theoretical quantum efficiency enhancement, , and are plotted in Figure 5 for a range of values. It is apparent that higher results in lower quantum efficiency enhancements (assuming the material has a similar refractive index to P3HT at 700 nm, i.e., ∼1.4) although modified quantum efficiency can remain large.28 Clearly the benefit of the split-dipole nanoantennas in terms of luminescence quantum efficiency applies to materials which possess relatively low values, but that may have other beneficial properties (such as high organic semiconductor hole mobility, in the case of P3HT). However, the radiative decay rate enhancement can remain high, regardless of the intrinsic quantum efficiency, therefore, where semiconductor radiative emission rate is the critical parameter, e.g., for optical communication or modulation applications, split-dipole plasmonic nanoantennas are attractive rate-enhancement tools. To conclude, this work has implications for a next-generation light-emitting and light-harvesting devices that could incorporate organic semiconductors with intrinsically low quantum efficiency, or for optoelectronic applications where the use of ultra-thin semiconductor material layers (i.e., ≤20 nm) would be beneficial. Dependence of modified quantum efficiency on intrinsic emitter quantum efficiency. Theoretical quantum efficiency enhancement, (squares) and modified quantum efficiency, , (circles) of a dipole emitter in the slot region of a split-dipole nanoantenna (L = 180 nm, d = 60 nm, nbk = 1.48 and free-space wavelength is 700 nm) as a function of intrinsic emitter quantum efficiency, . was calculated from the radiative and total decay rate simulated using 3-dimensional FDTD methods. Electromagnetic Simulations: Total decay rate was computed using 3D-FDTD software (Lumerical Solutions, Inc.) by the a surface only a dipole oriented parallel to the antenna long which was embedded in an constant to that of P3HT from and which was placed in the slot region of a gold split-dipole nanoantenna = 20 nm). The dipole nanoantenna the P3HT slot the dipole source were embedded in a with refractive of or 1.48. The radiative decay rate was computed by the a surface the non-radiative to the The non-radiative and radiative decay rate enhancements were calculated by to those of a dipole emitter in P3HT without the of the gold split-dipole nanoantenna. were nm for all simulations and, 2 nm for The intensity within the simulation was allowed to decay to an intensity of × of the source intensity has an intensity value of 1) simulations were to ensure that simulations had Nanoantenna Split-dipole nanoantennas were fabricated by template-directed sequential electrodeposition of gold and in nm nanoporous alumina and, subsequent thermal evaporation of gold into the alumina template The fabrication is as aluminum nm was thermally evaporated onto a conductive substrate or × was applied to the of the aluminum film to A of was to one side of the sample to as an to the aluminum during subsequent was applied to the on the of the aluminum film to and only the aluminum during The sample was in oxalic acid to in a with an aluminum A of 35 V was applied to the sample the aluminum was anodized by a clear change in to that of the conductive To the alumina barrier at the of the alumina template pores, the sample was in acid for 45 after which time the sample was in and under Gold was electrodeposited into the alumina from gold under a pulsed potential of -3 V for a given number of pulses and of 0.02 s duration V was applied for 5 s between using a in pulsed These conditions in gold segment lengths of 27 ± 6 nm per to regions with different gold segment lengths were on a single substrate by out each time the region that was to the the of the sample with a and the number of applied pulses. mesh was employed as both the electrode and the quasi-reference electrode during all gold deposition was the sample was from the gold and was with and and under For deposition of P3HT, monomer was at the of the electrodeposited gold segments under a pulsed potential of +3 V for 2 s duration with -0.6 V applied for 5 s between pulses to the polymer material and time for 3HT monomer to into the in time for the subsequent applied potential from a These conditions in P3HT segment lengths of nm per the alumina template was with to 3HT monomer), and and under to complete the nanoantennas, 40–80 nm of gold was deposited by thermal or evaporation at a rate of nm s−1 onto the the alumina template was the gold was deposited into the template pores onto the P3HT as well as on the surface of the alumina between The gold was from alumina surface by with a To the alumina the sample was in for and with and allowed to arrays of nanoantennas oriented vertically on the substrate. To nanoantennas, nanoantenna arrays were placed in an or for and was applied for to the nanoantennas in the P3HT nanowires were under the conditions as but without the gold electrodeposition or evaporation monomer nanoantennas were under similar conditions to the split-dipole nanoantennas but without the gold evaporation P3HT thin were by from a of the polymer Inc.) in at onto a The thickness of the P3HT film was nm by to optical measurements on nanoantenna a was deposited on the surface of an array to and of the semiconductor during For single nanoantenna nanowires were embedded in optical epoxy Inc.) and between Nanoantenna optical measurements were on an Inc.) with a of an 150 with and and a × For PL lifetime a nm ps duration depending on excitation source was used as it did not with surface plasmon resonances of the gold nanoantenna but could PL emission from P3HT. A ps nm, was also employed to excitation in the and wavelength range and the instrument response of the PL lifetime at different wavelength For the excitation a nm, Inc.) was used to of the to a A nm or a nm was placed after the at of the of the during PL measurements to the excitation at the sample to the of P3HT. excitation of × was The of the was to the of a single For nm the the to the The of a single active ps was used as the to the which was with a and with software The instrument response function of the PL lifetime was to have a of i.e., to the of the at nm using the of ps was at nm by the slightly response of the at PL lifetime decays were acquired by a followed by a a Inc.) to various regions of the sample and an the which was on an The was to light and detect PL from a single or a region of a nanoantenna array of than 5 in at the A was used to PL lifetime decays integrated the 650–750 nm wavelength range of P3HT The was to the of the for of nanoantennas in the A oil immersion objective of Inc.) was used for both scattered-light and PL For scattered-light the was light with full for plane excitation the nanowire sample of the A Inc.) was placed on top of the sample to (i.e., not lamp light to out of the sample reflection in to the objective at the top of the oil index of Inc.) was placed between the and the top substrate, between the and the sample and between the oil immersion objective and the of the is from the or from the This work was by the of in under and also by the of under support from the and for and of in the at of is and with and during the of this work are of to are as are but not or are as by the The is not for the or of by the than be to the for the